7/26/2019 Vede Js 1995

1/8

3020 J . Org. Chem. 1995,60, 3020-3027

Conversion of Arylboronic Acids into PotassiumAryltrifluoroborates: Convenient Precursors of Arylboron

Difluoride Lewis Acids

E. Vedejs,* R. W. Chapman, S. C. Fields, S. Lin, an d M. R. SchrimpfChemistry D e p a r t m e n t , U n i v e r s it yof Wisconsin,Madison, Wisconsin 53706

Rece ived Sep tember30 1994 Revised Manuscr ip t Rece ived March13, 1995@)

Reaction of ArB(OH)2 (3) with KHF2 affo rds crys tall ine s al ts JSArBF3 2). n the p resence of TMSClin acetonitrile, 2a eacts to give NMR s ignals typ ical of PhBFz in acetonitrile solution. When thereaction of 2 + TMSCl is performed in th e presence of potential Lewis bases, the trivale nt borane1 is intercepted, resul ting i n organoboron complexes. Thus, th e oxazaborolidinones 7- 10 havebeen prep ared from amin o acid-derived amidin e carboxylat es NaOZCCH(R)N=CHNMez (R = H orphenyl). Complexes 11 and 12 derived from 1,3-diketones are also easily prepared. The KHFzfluoride exchange coupled with th e TMSCl activation method allows in situ generation of ArBFzwithout having to handle corrosive trivalent boron halides.

We have been interest ed in th e preparation of aminoacid-derived oxazaborolidinones as par t of our study ofcrystallization-induced asymmetric transf0rmation.l In

the early stages of this work, we prepared severaloxazaborolidinones from amino acid derivati ves by tre at-ment with PhBFz The trivalent boron reagent wasprepa red from PhBClz according to the published method:but both the starting material and the product wereunpleasant to handle. Similar problems have beenencountered with other arylboron di fl uo ri de ~. ~~ ~ ince weanticipated the need to explore several different aryl-boron environments, experiments were initiated to finda convenient way t o generate ArBFz in situ ideallywithout using corrosive trivalent boron reagents.

Several strategies have been reported in the literat ureth at circumvent the problem of handli ng reactive Lewisacids. One option is to use the relatively stable Lewisbase complexes (e.g., BH3aSMe2, RZHB-MezNCHzCHz-NMez*BHR2, iClgAsPh3) as precursors,4 and another isto employ an ionic ate species (sa lts such as KRBF3,LiBF4, NaBP h4, LiMeBH3) as th e s tar tin g material^.^,^ ^Under suitable activation conditions, he tetravalent atecomplexes may function as in situ sources of the trivalentborane derivatives. This is illustrate d by the reversibledissociation of LiBF4 into BF3 LiF a t room tem-p e r a t ~ r e , ~ ~y the acid treatm ent of LiMeBH3 to generateM ~ B H z , ~ ~r by the conversion of RzHB-MezNCHzCH2-N M e y B m into &BH upon treatment with BF3.etherate.4CIn a more subtle example, NaBPh4 can be used as areplacement for PhzBX reagents in the conversion of

Abstract published in Advance ACS Abs t rac ts ,May 1, 1995.1) a) Vedejs, E.; Fields, S. C.; Schrimpf, M. R. J. m. Chem. SOC.

1993,115, 11612. b)Vedejs, E.; Fields, S. C.; Lin, S.; Schrimpf, M. R.J . Org. Chem. 1994, 59, 0000 accompanying Ms.) c) The X-raystructures of a ser ies of related oxazaborolidinones will be publishedin due course: Powell, D. R.; Chapman, R. W.; Fields, S. C.; Lin, S.;Schrimpf, M. R. Manuscript in preparation.

(2 )McCusker, P. A,; Makowski, H. S. J . Am. Chem. SOC. 957, 79,5185.

3) Bir, G.; Schacht, W.; Kaufmann , D. J . Organomet. Chem . 1988,340, 267.

4) (a) Braun, L. M.; Braun, R. A.; Crissman, H. R.; Opperman, M.;Adams, R. M. J . Org. Chem. 1971,36,2388. b) Suzuki, I.; Yamamoto,Y. J . Org. Chem. 1993,58,4783. c) Ganesan, K.; Brown, H. C. J . Org.Chem. 1994 ,59, 2336.

(5 ) a) Greenwood, N. N. Quar t . Reu. 1954,8, 1. See also Smith, D.A,; Houk, K. N. Tetrahedron Lett 1991, 32, 1549. b) Cole, T. E.;Bakshi, R. K.; Srebnik, M.; Singaram, B.; Brown, H. C. Organometallics1986, 5, 2303.

6)Ba um, G. J . Orgunomet. Chem. 1970,22, 269.

0022-3263/95/1960-3020 09.00/0

amino acids into the crystalline and easily isolated 2,2-diphenyl-1,3,2-oxazaborolidin-5-ones, erivatives that canbe convenient for purposes of purification. Presumably,

hydrolytic cleavage converts the P b B - ion into a morereactive trivalent intermediate in this case.6In the most relevant prior example, Kaufmann et al.

showed by llB NMR t ha t K(ipc)BF3 libe rate s (ipc)BFzwhen t reat ed w ith BF3.OEtz in acetoni trile (ipc = isopi-n ~ c a m p h e y l ) . ~ his experiment demonstrates the pos-sibility of releasing an unstable chiral Lewis acid from arelatively stable alkyltrifluoroborate salt. A literaturesearch also uncovered isolated reports mentioning potas-sium aryltrifluoroborates7 and vinyltrifluoroborates,8.9structures that might serve as suitable precursors ofsubstituted difluoroboranes. For instance, Stafford pre-pared K(CHz=CH)BF3 from (CH2 =CH )B F~ nd KF anddemonstrated that it reverts t o the starting materialswhen heated to 250 oC.s In contrast t o the trivalentvinyldifluoroborane, which hydrolyzes readily t o ethyl-ene, the tetravalent ate salt potassium vinyltrifluo-roborate is stable for days in water. Despite theseimportant observations, the synthetic potential of thealkyl- or aryltrifluoroborate species as in situ sour ces oftrivalen t boron halide Lewis acids has not attract ed muchattention. In this paper we report an expedient synthesisof a number of stable salts KArBF3 2) nd their facileactivation by fluorophiles to give intermediates thatfunction as sources of the corresponding aryldifluorobo-ranes 1. In addition, we demonstrate the use of 2 orthe preparation of boron heterocycles from amino acidderivatives or 1,3-dicarbonyl compounds.

For our initial studies on the reactivity of the boratesalts , KPhBF3 2a) as prepared from PhBClz and KF_ _ _ _

7) (a) Thierig, D.; Umland, F. Naturwissenschaften 1967,54, 563.b) Chambers, R. D.; Chivers, T . J . Chem. SOC. 1965, 3933. c )

Chambers, R. D.; Chivers, T.; Fyke, D. A. J . Chem. SOC. 965, 5144.d) Chivers, T. Can. J. Chem. 1970,48,3856. e) Fowler, D. L.; Kraus,

C. A. J . Am. Chem. SOC. 1940, 62, 1143.(8) Stafford, S. L. Can. J Chem. 1963,41, 807. Chem. Abs t r. 1965,

62, 9173.9) (a) Chambers, R. D.; Clark, H. C.; Willis, C. J. J . Am. Chem.

SOC. 960, 82, 5298. b) Chambers, R. D.; Clark, H. C.; Willis, C. J.Proc. Chem. SOC. 960, 114. c ) Booth, M. R.; Cardin, D. J.; Carey, N.A . D.; Clark, H. C.; Sreenathan , B. R. J . Organometal. Chem . 1970,21, 171. d) Jander, J.; Nagel, H. Liebigs Ann. Chem.1963, 669, 1. e)Pawelke, G.; Heyder, F.; Burger, H. J . Organometal. Chem . 1979,178,1. 0 Brauer, D. J.; Burger, H.; Pawelke, G. J Organometal . Chem.1980,192, 305.

0 1995 American Chemical Society

7/26/2019 Vede Js 1995

2/8

Synthesis of Potassium Aryltrifluoroborates

Scheme 1

I

Ar r

1 2

a Ar=C6H5

X

d

Xc l wh

?rAr Ar

3 4

b Ar= l-Naphthyl

X x

e f

O X

i

1 X = B F p ; 2 X=BF3 ; 3 X=B[OHJ2 ; 4 X = B r ;

by analogy to the procedure of Kaufmann et al.3 Ulti-mately, however, this approach is not ideal because itstill requires dealing directly with the reactive RBXzstarting material. A far more convenient alternativewould be t o prepare the desired salts directly fromboronic acid precursors th at are readily available and air-stable. The potential advantages of this appro ach wererecognized early on,2 but were frustrated by the sensitiv-ity of arylboronic acids t o protodeboronation. A possiblesolution t o the problem was deduced from a literatureanalogy.7a Thus, Thierig and Umland noted that the

reaction of the ethanolamine complex of PhzBOH (2,2-diphenyl-1,3,2-oxazaborolidine) ith aqueous KHFz re-sults in formation of KPhZBFz. This obser vatio n showsth at t he inexpensive KHFz can function as a flu oride ionsource, and t hat it can activate a relatively unreactiveboronate structure for ligand exchange under weaklyacidic conditions. It also suggests th at the diphenyldi-fluoroborate anion may be therm odynam ically morestable under the reaction conditions than is PhzBOH(diphenylborinic acid), th e initial product expected fromate dissociation and hydrolytic cleavage. Thierig andUmland noted that heating th e same reactants in glacialacetic acid gave KPhBF3 (2a) n unstated yield. Whilethey did not provide experimental details or discuss th emechani sm, this transformation would appear t o involveintermediates having one B-phenyl and a t least oneB-oxygen bond, and might well involve phenylboronicacid, the most desirable starting material for directconversion t o the potassium phenyltrifluoroborate salt.A simple, high-yielding procedure for th e synthesis of 2awas devised by extrapolating from these ob servations asfollows. Treatment of a concentrated solution of phenyl-boronic acid in methanol with satur ated aqueous KHFzresulted in an exothermic reaction and immediate forma-tion of a precipitate. Collection of the cr ystals by filtra-tion and recrystallization from acetonitrile afforded an82% yield of spectroscopically pure KPhBF3 (2a). Thefluoride exchange did not ta ke place when K F was usedin place of KHFZ. The phenyltrifluorobo rate salt obtained

J. Org. Chem., Vol 60, No. 10, 1995 3021

Li HAr Ar

5 6

c Ar= pCF3C6H4

X

9

i

5 X = L i : 6 X = H

as described above was not appreciably hygroscopic andcould be stored in air for many mo nths witho ut signifi-cant decomposition. The structu ral assignme nt wasconfirmed by the llB NMR spectrum, which has achemical shift appropriate for tetracoordinate boron (64.1 ppm relati ve to borontrifluoride ethe rate)1 and showscoupling t o three e quivalent fluorine atoms = 47 Hz).Aqueous solutions of the salt are somewhat acidic,suggesting the existence of an equilibrium between thetetracoordinated ate species and products of hydrolyticcleavage or ligand exchange. The efficient formation of

the salt in aqueous methanol must therefore be aconsequence of the the rmodyn amic stability of the phen-yltrifluoroborate anion and of the added driving forceprovided by it s low solubility in th e reaction m edium.

A number of other aryltrifluoroborates have beenprepared by this method (Table 1; Scheme 1). Therequisite arylboronic acid precurs ors 3b-d were synthe-sized using varia nts of the method of Snieckus et al. andBrown et al.llayc Thus, commercially available arylbromides 4b-d were converted into the aryllithiumreagents 5b-d by exchange with n-butyllithium or tert-butyllithium (Table 1; method A). Reaction with B(OiPrI3or B(OMe)3 and hydrolysis then gave the arylboronicacids 3b-d. In the other examples (Table 1; method B),the aryllithium reagents were prepared from 6e-j bydirect ortho-metalation of the activated aryl C-H bondsusing n-butyllithium (5h,i) r sec-butyllithium Se-gj).The metalation approach t o arylboronic acids has beenused extensively, although oth er activating groups wereusually involved.la As before, reaction of 5e-j withB( OiPr)3 or B(OMeI3 followed by hydro lysis gave arylbo-ronic acids. The crude arylboronic acids were typicallyused without purification for the conversion to potassium

(10) Noth, H.; Wrackmeyer, B. In Nuclear M agnetic ResonanceofBoron Compounds;Springer-Verlag: New York, 1978.

11) a)Alo, B. I.; Kandil, A.; Patil, P. A,; Sharp, M. J.; Siddiqui, M.A,; Snieckus, V.; Josephy, P. D. J. Org. Chem 1991, 56, 3763. b)Thompson, W. J. ; Gaudino, J . J. Org. Chem. 1984,49,5237. c) Brown,H. C.; Cole, T. E. Organometallics 1983,2,1316. Brown, H. C.; Srebnik,M.; Cole, T. E. Organometallics 1986, 5, 2300.

7/26/2019 Vede Js 1995

3/8

3022 J Org Chem., Vol. 60, No. 10, 1995

Table 1. Preparation of Aryltrifluoroborates 2

star ting yield yieldentry material method RLij o f 3 KArBF3 o f 2

2a 80PhB OHj22 4b A n-BuLij 84a 2b 823 4c A tert-BuLi) b 2c 9 f4 4d A tert-BuLi) b,c 2d 68f5 6e B sec-BuLij 81 2e 946 6f B sec-BuLij 92 2f 937 6g B sec-BuLi) 93 2g 768 6h B n-BuLi) b,d 2h 539 6i B n-BuLi) b,e 2i 48f

10 6j B sec-BuLij b 2j 7Gg

a 3b: ref 15b. The arylboronic acid 3 was not purified; one-pot conversion into 2. 3d: ref 15a. 3h: ref 15a. e 3i: ref 16.f Overall yield based on the indicated starting mate rial far leflcolumn). g The sta rting menthy l p-fluorophenyl ether was madeby Ullm an coupling ref 1 7 ) from p-fluorobromobenzene.

aryltrifluoroborate salts 2e-j. This is feasible becausebyproducts such a s trimeric or oligomeric boronic anhy-drides also appear to be reactive in the KHFz fluorideexchange procedure. Thus, several aryltrifluoroborateswere obtained in good yield using a one-pot approach(Table 1; entries 3, 4, 11). Lower yields in the furanexample reflect increased difficulty in th e metal ationlZ

and grea ter sensitivity of th e 2-furylboronic acids to acid-induced protodeboronation.The best re sults were obtained wi th th e fluorophenyl

or (trifluoromethy1)phenyl substra tes. The correspondingfluorine-substituted arylboronic acids are resistant toprotodeboronation, and 3e-g are also relatively well-behaved i n the s ense tha t the y can be crystallized withrelatively minor contamination by boronic anhydrides.However, th e relative puri ty of the fluoroaromatic boronicacids is not the decisive factor in the yield of 2. Asalready noted, even a relatively stable ethanolamine-derived ate complex of diphenylborinic acid i s reactivewith K H F z . ~ ~ hus, it seems likely that any aryl-substituted su bstrat e having heteroatom bonds a t tet-ravalent o r trivalent boron would be subject to the sameconversion into pot assium fluoroborate s alt s 2, rovidedth at protodeboronation is controlled.

In all of the fluoroaromatic substrates, metalationoccurred selectively next to th e fluorine substitu ent usingsec-BuLi in THF at -78 OC.I3 T his re su lt was easilypredictable for 6e-g, but i t was not clear tha t the s amepreference should be expected in the case of phenylmenthyl ether 6j (the precursor to 5j and 29. Goodselectivity is somewhat surp risi ng in view of prior experi-ence with t he me tala tion of p-alkoxyfluorobenzene de-r i v a t i v e ~ . ~ ~n th e absence of strong lit hium complexingagents, metalation next to alkoxy oxygen can be competi-tive or even dominant. However, there is some evidencethat the relatively bulky TBS ethers are not effective

directing groups in these meta1 ati 0ns .l~ ~ pparently, t hementhyloxy substitue nt is also too hindered to promoteortho metalation , and the fluorine directing effect domi-nates.

The isolated yields of aryltrifluoroborate sa lt s wereusually in t he range of 70 or better, but th ere were twoexceptions. In the fura n example (entr y 9)) yields werelow because the inter media te boronic acid 3i is sensitiveto protodeboronation. A different problem was encoun-

Vedejs et al.

tered in the boronation of (2,6-dichlorophenyl)lithiumcase, probably because 5h decomposes at -50 C orabove.14 However, th e corresponding sal t 2h was stableand easy to handle.

With a variety of aryltrifluoroborates available, atten-tion was turned to t he use of these reagents for the insitu gene rati on of arylboron difluorides. This conversionrequires an agent that can assist in the removal offluoride from tetrahed ral boron. Previous repor ts indi-cate that the pa rent tetrafluoroborate salt s Mf BF4- vary

greatly in stability, depending on th e cation. Thus, LiBF4is in equilibrium with BF3 well below room tempe rat urewhile NaBF4 begins to decompose only above 270 C andKBF4 survives to even hi gher temperature^.^^ In addi-tion to th e difference in kinetic stability, there is a largedifference in dissociation enthalpy that strongly favorsKBF4 over K F BF3, in contrast to the behavior ofLiBF4.5a Not surprisingly, i t was found tha t t he stabilityof the aryltrifluoroborate ion is also highly depe nden t onthe identity of th e metal counterion. Several potentialfluorophiles)) were screened usin g llB NMR spectroscopyto monitor the fate of the PhBF3- anion. It was foundth at KPhBF3 rapidly decomposes to phenylboronic acidin the presence of lithium or magnesium cations, pre-

sumably via dissociation into PhBFz and t he lithi um ormagnesi um fluorides followed by hydrolysis. This resu ltis consistent wit h th e well-known fact th at t he additionof organolithium or Grignard reag ents to BF3 resul ts inmultiple additions of the organometallic reagent.2 Thisoccurs because the intermediate ate species can easilydissociate t o give LiF or MgFz, thereby regeneratingreactive trivale nt boranes. Thus, Mgz+ or Li couldpossibly serve as fluorophiles that convert aryltrifluo-roborate species into th e arylboron difluorides. However,the complication of having to use anhydrous lithium ormagnesium salts to avoid B-F bond hydrolysis promptedus to look for more convenient fluorophiles. We turne dour attention to silicon-containing compounds.

The st rengt h of the Si-F bond is the driving force fora num ber of common reactions, such a s the deprotectionof trialkylsilyl ethe rs and t he generation of enolates fromenol silanes with tetrabutylammonium fluoride. In thepresent context it is significant t ha t LiBF4 can be usedas a reagent for silyl ether dep r~te ction ,~ transforma-tion th at appe ars to involve th e gene rati on of BFa, Lewisacid activation of oxygen, an d ultimately, th e tran sfe r offluoride from the ate complex to silicon. In effect, th esilyl ether acts a s a fluorophile that drives the kineticallyfacile dissociation of lith ium tet rafl uoro bora te anion. Asexpected from this analogy, the relatively electron-deficient trimethylsilyl chloride was found to be aneffective fluorophile in t he reaction with aryltrifluorobo-rates . This simple activation technique h as proven to

be suitable for the generation and trapping of specieshavi ng th e reactivity expected for ArBF2.The llB resonance of 2a n acetonitrile (6 4.1 ppm) is

a qu art et due to coupling to three fluorine substitu ents.

(12) y, N .; chlosser, M. Helu. Chim. Acta 1977, 0, 2085.(13) a) Bridges, A. .; Lee, A.; Maduakor, E. C.; Schwartz, C. E .

Tetrahedron Lett. 1992,33, 495. b) Furlano, D. .; Calderon, S. N.;Chen, G . ;Kirk, K. L. J. Org. Chem. 1988,53, 145. c) Katsoulos, G.;Takagishi, S.; Schlosser, M. Synlett 1991, 31.

(14) a) Kress, T. .; Leanna, M. R. Synthesis 1988, 03. b) Iwao,M. J . Org. Chem. 1990,55,3622.

(15) a) Poole, C. F.; Singhawangcha, S.; Zlatkis, A. . Chromatog.1979, 86, 307. b) Poole, C. F.; Singhawangcha, S.; Zlatkis, A. .Chromatog. 1978, 58, 33.

(16) oques, B. P.; Florentin, D.; Callanquin, M. J . Heterocycl. Chem.1975, 2, 95.

(17) Whitesides, G. M.; Sadowski, J. S.; Lilburn, J. J . Am. Chem.SOC.1974, 6, 2829.

(18) Metcalf, B. W.; urkhart, J. P.; Jund, K. Tetrahedron Lett. 1980,21, 35. Lipshu tz, B. H.; egram, J. J.; Morey, M . C. Tetrahedron Lett.1981,22,4603.

7/26/2019 Vede Js 1995

4/8

Synthesis of Potassium Aryltrifluoroborates

Scheme 2

J. Org. Chem., Vol. 60, No. 10 1995 3023

7-trans-E

8-trans-EAF GH5

Ar= 2-FGH4

KArBF3

Me3SiCl

KArBF3

c, 5U G H 5

When a trace of TMSCl(

7/26/2019 Vede Js 1995

5/8

3024 J . Org. Chem., Vol. 60 No. 10 1995

reached at room tempera ture. Both the 6 5.21 a nd 5.45ppm signals were observed as doublets due to fluorinecoupling 4J = 3.7 Hz). Fur the r warming in dichloro-methane resulted in the gradual appearance of signalsassigned to 7-cis, but the tran dcis equilibration was moreconveniently monitore d in CD3CN. Afte r warmi ng to 70C in th is solvent, two new signals for 7-cis-E and 7-cis-Z

appeared (6 5.68, 5.28 ppm; 1:l ratio) in addition to themethine signals of 7-trans-E 6 5.33 ppm) and 7-trans-Z(6 5.62 ppm). In contra st to 7-trans, he methine signalsof 7-cis were not sp lit appreciably by fluorine (singlets;4J = < 1 Hz). Attempts to purify 7-cis invariably gaveisomer mixtures due to the facile interconversion ofdiastereomers.

As already mentioned, the ratio of 7-trans:7-cis atequilibrium was 3:l at 70 C. Surprisingly, this productratio was considerably altered (as high as 99:l 7-trans:7-cis) after removal of solvent on a rotary evaporator(bath temperature 30-40 C). This behavior is due tothe crystallization-induced asymmetric transformationphenomenon discussed in the accompanying paperlb andinvolves reversible epimerization at boron. The alterna-tive possibility of epimerization at carbon was ruled outby cleavage of 7-trans to t he sta rti ng (R)-phenylglycine.

Thus, 7-trans was dissolved in warm methanol andethylenediamine (5 equiv), and a cataly tic amo unt of HC1was added. &r solvent re moval, (R)-phenylglycine wasrecovered (88%) with no change in optical rota tion ('95%ee). This evidence confirms boron epimerization as t hemechanism for interconversion of 7-cis and 7-trans andalso demonstrates th at synthesis of 7 using the in situgenera tion of PhBFz proceeds wit hout racemiz ation of thephenylglycine substrate.

Crystallization-induced asymmetric transforma tion wasnot encountered in the B-(2-fluorophenyl) series 8)derived from le. Conventional chromatography or crys-tallization could be used to sep arat e the relatively stablediastereomers 8-trans and 8-cis. In contrast to theB-phenyl analog 7-trans, he o-fluorophenyl derivative8-trans did not isomerize, even after he ating to 100 Cin toluene. Traces of equilibration did occur when8-trans or 8 4 s were heated in acetonitrile, and somedecomposition was also observed under t hese conditions.Attempts to recrystallize 8-cis may also have encoun-tered minor equilibration, and this isomer could not beobtained completely pure (ca. 5% of 8-trans contami-nant ). However, th e interconversion of 8-cis and 8-transwas considerably slower th an the analogous process inthe B-phenyl series (7). This observation is consistentwith reversible B-N cleavage as the mechanism forboron epimerization.lb The relatively electronegativefluorophenyl group stabilizes the ate complex 8 andpreven ts boron epimerization via trivalent intermediates.

Conversion of 1,3-dicarbonyl compounds into cyclicboron complexes was also examined briefly using thesame TMSCl activation method. Thus, l-phenylpentane-1,3-dione was sti rred with t he s alts 2e or 2.j in THF atroom tem per atur e with excess TMSC1. Stru ctur e 12 wasobtained as an inseparable m ixture of two diastereomers(two epimers at stereogenic boron relative to stereogeniccarbon in the menthyloxy substituent) and did notcrystallize. However, no geometrical isomers are possiblewith 11. This substance crystallized upon solvent re-moval an d was obtained in nearly quantitative yield.

In summary, we have prepared a variety of KArBF3salts 2. The salt s ar e available on multigram scale fromarylboronic acids and KHFz. This method solves an old

Vedejs et al.

problem by providin g access to reac tive arylbor on difluo-rides without resorting to corrosive reagen ts. The potas-sium aryltrifluoroborate salts 2 are crystalline, water-resistant materials that can be stored without specialprecautions. Generat ion of reactive Lewis acids occursunder mild conditions upon treatment with chlorotri-methylsilane as the fluorophile, and conversion intoboron-containing heterocycles is possible i n th e presenceof difunctional reactants.

Experimental Section

Potassium Phenyltrifluoroborate (2a). Phenylboronicacid (20 g, ca. 169 mmol, Aldrich; ca. 80% PhB(OH)Z, ca. 20%(PhBO),) was dissolved in 50 mL of methanol. Excess satu-rated KHFz (125 mL, ca. 4.5 M solution, ca. 563 mmol) wasadded slowly with vigorous stirring. After 15 min, theprecipitated product was collected and washed with coldmethanol. Recrystallization from minimal acetonitrile pro-duced 25.5 g (138 mmol, 82%) of pure 2a, mp 296 "C dec, lit.290 0C);7a nal. calcd: C, 39.16; H, 2.74; found: C, 39.12; H,3.02; 200 MHz NMR (CDaCN, ppm) 6 7.44-7.41 (2H, m), 7.22-7.05 (3H, m); 160 MHz IlB NMR (CDsCN, ppm) 6 4.1 (9, J =57 Hz); 470 MHz 19F NMR (CDsCN, ppm vs CF3CsHd 6 -79

1:l:l : l , J = 57 Hz).Potassium 1-Napthyltrifluoroborate 2b). The proce-

dure was similar to tha t described for 2a. Pure material (0.221g, 82% from 0.2 g of 1-naphthylboronic acid15b) was obtainedby extracting the initial precipitate with hot acetonitrile (2 x10 mL), evaporation, and recrystallization of the residue fromhot acetonitr ile, mp 205 "C dec, anal. calcd: C, 51.31; H, 3.01;found: C, 50.99; H, 3.16; 200 MHz NMR (CD3CN, ppm) 68.42-8.39 ( lH , m), 7.78-7.71 (l H, m), 7.63 (2H, d, J = 7.7Hz), 7.40-7.29 (3H, m); 160 MHz IlB NMR (CD3CN, ppm) 64.4 (9, J = 54 Hz).

Potassium 3,5-Bis(trifluoromethyl)phenyltrifluoro-borate (2d) by the One-Pot Procedure (Method A). To asolution of 3,5-bis(trifluoromethyl)bromobenzene 4.0 mL, 23.0mmol) in 20 mL of ether a t -78 C was added dropwise 2 equivof tert-BuLi (1.7 M in pentane, 27 mL, 46.0 mmol). Theresulting solution was stirred a t -78 "C for 2 h and another15 min at room temperature. The lithium reagent thus

obtained was chilled with a dry ice-acetone bath and droppedinto a solution of B(Oi-Pr)a 5.3 mL, 23.0 mmol, distilled fromsodium) in 100 mL of ether via cannula at -78 "C. Afterstirring for another 2 h at this temperature, the reactionmixture was allowed to warm to 0 "C and was then quenchedby addition of 30 mL of HzO. After the pH of the aqueouslayer was adjusted t o 2 with concentrated HC1, the organicphase was separated, and the aqueous phase was extractedwith ether (2 x 30 mL). The organic solution was combined,dried (MgS04), and concentrated to give an oil containing theknown boronic boronic anhydrides, and residual bo-ronic esters due t o incomplete hydrolysis. This mixture wassatisfactory for the conversion to the potassium aryltrifluo-roborate salt in all examples where the arylboronic acid wasnot purified.

The crude boronic acid obtained above was refluxed with

4.9 g of KHFz in 100 mL of MeOH and 20 mL of HzO for 12 h.The solution was concentrated (rotary evaporator), and thesolid residue was extracted with CH3CN (3 x 20 mL, roomtemperature) . After filtrat ion and evaporation of the aceto-nitrile, pure material (5.0 g, 68% from 3,5-bis[trifluoromethyll-bromobenzene) was obtained by recrystallization from ether/hexane, mp 320 "C dec, colorless crystals. Anal. Calcd: C,30.02; H, 0.94, found: C, 29.70; H, 0.63; IR (KBr, cm-') 1619,C=C; 1128, C-F; 200 MHz NMR (CD3CN, ppm) 6 7.96 (2H,s) 7.72 (lH, s).

Potassium 4-(Trifluoromethyl)phenyltrifluoroborate(2c). Method A was used to prepare the arylboronic acid from4-(trifluoromethy1)bromobenzene Aldrich, 3.2 mL, 22.8 mmol),tert-BuLi (26.9 mL, 1.7M in pentane, 45.6 mmol), and B(OiPrh(5.3 mL, 22.8 mmol). After the usual workup, the crudearylboronic acid was dissolved in 100 mL of methanol and 20

7/26/2019 Vede Js 1995

6/8

Synthesis of Potassium Aryltrifluoroborates

mL of water, and 5.9 g of KHF2 (Aldrich, 75.6 mmol) wasadded. After refluxing for 36 h, the reaction mixture wasconcentrated to dryness by aspirator, and the solid residue wasextracted with hot acetonitrile (2 x 30 mL). Then thecombined acetonitrile solution was concentrated t o dryness,and the solid was washed with ether to give essentially pureproduct (5.3 g, 91% from 4-(trifluoromethy1)bromobenzene).Recrystallization from ethyl acetabdether gave 2d, mp 305 Cdec, as colorless crystals. Anal. Calcd: C, 33.36; H, 1.60,found: C, 32.97; H, 1.33; IR (KBr, cm-') 3091, =C-H; 1331,C-F; 959, B-F; 200 MHz NMR (CD&N, ppm) 6 7.60 (2H, d,

CN, ppm) 6 3.3.2-Fluoro-5-methylphenylboronic cid (30 y Metala-

tion of 6f (Method B). A solution of 4-fluorotoluene (5.0 mL,45 mmol, Aldrich) in 100 mL of dry THF was chilled in a dry-ice/acetone bath under nitrogen. sec-BuLi (48 mL of a 1.0 Msolution in cyclohexane, 48 mmol, Aldrich) was added over 10min, and the resulting yellow solution was allowed t o sti r foran additional 10 min. Trimethyl borate (Aldrich, 5.4 mL, 48mmol, distilled over sodium) was added over 1 min, and thesolution was allowed t o warm to room temperature. Thereaction was quenched by the additon of 50 mL of 10% aqueousHC1, and the mixture was diluted with 50 mL of ether. Theorganic portion was saved and was extracted with 1 N NaOH(2 x 50 mL). The basic extracts were combined an d acidifiedto pH 3 by the addition of 10% aqueous HC1. The mixturewas ether extracted (3 x 50 mL), and the organic layer wassaved and subsequently concentrated (aspirator) to yieldcolorless crystals (6.7 g, 96%). Analytical TLC on silica gel,EtOAc, Rf 0.72. Pure material (6.4 g, 92%) was obtainedby crystallization from dichloromethane, mp 168- 170 C dec.Due to the ease of boronic anhydride formation, this materialwas analyzed after the next step, at the stage of the salt 2f.Data for 3f: IR (KBr, cm-l) 3300,O-H; 1400, B-0; 200 MHzNMR (CDC13, ppm) 6 7.47 (l H, dd, J = 2.1, 6.2 Hz), 7.25 ( lH ,ddd,J=2.1,5.6,8.6Hz),6.94(1H,dd,J=8.6,9.9Hz),6.12( lH , s), 6.11 (lH, s), 2.3 (3H, d, J = 0.5 Hz).

2-Fluorophenylboronic cid (3e). Method B was usedstarting with fluorobenzene (5.0 mL, 53 mmol, Aldrich), sec-BuLi (56 mL of a 1.OM solution in cyclohexane, 56 mmol,Aldrich), and trimethyl borate (6.6 mL, 59 mmol). After theusual workup, the mixture was ether extracted (3 x 50 mL).The organic solution was allowed t o sta nd, yielding 4.2 g ofcolorless plates. An additional crop yielded 1.84 g of colorlessplates (6.04 g total, 81%). Analytical TLC on silica gel, EtOAc,Rf 0 77 Nearly pure material was obtained by crystalliza-tion from ether , mp 215-217 C dec. Due to the ease of boronicanhydride formation, this material was analyzed after the nextstep, at the stage of the sa lt 2e. Data for 3e: IR (KBr, cm-l)3355, 0-H; 1381, B-0; 200 MHz NMR (CDC13, ppm) 6 7.85(l H, ddd, J = 1.8, 7.1, 7.3 Hz), 7.45 (l H, dddd, J = 1.8, 6.2,7.3, 8.2 Hz), 7.20 ( lH, dddd, J = 0.8, 0.8, 7.3, 7.3 Hz), 7 05(lH , ddd, J = 0.8, 8.2, 10.8 Hz), 5.86 (l H , s) 5.82 (lH, s).

Potassium (2-Fluorophenyl)trifluoroborate 2e). To asolution of (2-fluoropheny1)boronic acid (2.0 g, 14.4 mmol) in20 mL of methanol was added aqueous KHFz (14.4 mL of a3.OM solution in water, 43 mmol). The resulting precipitatewas stirred for 20 min, and the solid mass was filtered. Thesolid was dissolved in 20 mL of hot acetonitrile, filtered, and

allowed to stand . Colorless crystals (0.98 g) were filtered, andthe mother liquor was allowed to stand to yield an addtionalcrop (1.23 g, 2.73 g overall or 94%). Analytical TLC on silicagel, EtOAc, Rf = 0.32. Pure material was obtained bycrystallization from acetonitrile, mp 304-305 C, colorlessplates. Anal. Calcd: C, 35.67; H, 2.00, found: C, 35.38; H,2.11; IR (KBr, cm-l) 3080, =C-H; 1189, B-F; 200 MHz NMR(CD&N, ppm) 6 7.55-7.42 (l H, m), 7.07 (l H, dddd, J = 7.6,

m); I1B NMR (160 MHz, CD$CN, ppm) 6 3.42 (9, J = 49 Hz).Potassium (2-Fluoro-5-methylpheny1)trifluoroborate

(20. To a solution of (2-fluoro-5-methylpheny1)boronic cid(5.0 g, 33 mmol) from method B, above, in 40 mL of methanolwas added aqueous KHF2 (33 mL of a 3.0 M solution, 99 mmol).The resulting suspension was stirred for 20 min, and removalof solvent (aspira tor) provided a solid mass which was dis-

J = 7.8 Hz), 7.46 (2H, d, J = 7.8 Hz); 160 MHz B NMR (CD3-

7.6,6.6,2.0H~),6.91(1H,dd,J=6.8,6.6H~),6.80-6.70(1H,

J. Org. Chem., Vol. 60 No. 10 1995 3025

solved in hot acetonitrile and suction filtered. The solutionwas concentrated (aspi rator) o yield crystalline materia l (6.9g, 97%). Analytical TLC on silica gel, EtOAc, R f = .32. Purematerial (6.6 g, 93%) was obtained by crystallization fromacetonitrile, mp 265-267 C as colorless needles. Anal.Calcd: C, 38.91; H, 2.81, found: C, 38.66; H, 3.04; IR (KBr,cm-') 3042, =C-H; 990, B-F; 200 MHz NMR (CDsCN, ppm)6 7.16 (lH , d, J = 2.9 Hz), 6.87 ( lH, ddd, J = 2.9,5.3,8.7 Hz),6.66 ( lH , dd, J = 8.7, 8.7 Hz), 2.18 (3H, d, J = 0.8 Hz); llBNMR (160 MHz, CD&N, ppm) 6 3.32 (9, J = 49 Hz).

Preparation of 1 -@-Fluorophenyl)-2-methyl-l-pro-

pene (6g). Step 1. l-@-Fluorophenyl)-2-methyl-l-pro-panol. A solution of 4-fluorobenzaldehyde (Aldrich, 20 mL,186 mmol) in 100 mL of anhydrous ether was chilled in anice-water bath and to it was added dropwise isopropylmag-nesium chloride (Aldrich, 120 mL of a 2 M solution in ether,240 mmol, 1.3 equiv). The resulting suspension was sti rredfor 1 h and was allowed to warm to room temperature over 1h. The suspension was rechilled in an ice-water bath , and100 mL of saturated aqueous ammonium chloride was addedslowly. The organic layer was saved and washed with distilledwater (2 x 100 mL) and brine (2 x 100 mL). Removal ofsolvent (aspirator) yielded a liquid (29.1 g, 93% crude), whichwas distilled (0.5 torr, 63-65 C) t o provide 28.7 g of the tit lecompound (92% yield), which was sufficiently pure for the nextstep.

Step 2. l-@-Fluorophenyl)-2-methyl-l-propene 6g).A solution of 1-(p-fluorophenyl)-2-methyl-l-propanol(5.1 , 30mmol) and p-TsOH-HzO 0.57 g, 3 mmol) in 80 mL of toluenewas refluxed under a Dean-Stark trap for 40 min. Aftercooling, saturated aqueous sodium bicarbonate (30 mL) wasadded, the organic layer was extracted with water (2 x 25 mL)and brine (2 x 25 mL) and was dried over NaZS04. The solventwas removed (aspirator) to give 4.3 g of an oil (96%), and shortpath distillation gave a clear liquid, bp 183-185 C (740 mm);analytical TLC on silica gel, hexane, Rf 0.46; molecular ioncalcd for CloH11F: 150.08447; found m e = 150.0848, error =2 ppm; base peak = 135 amu; IR (neat, cm-l) 2970, =C-H,1506, C=C; 1227, C-F; 200 MHz NMR (CDC13, ppm) unknownminor impurity, singlets at 6 4.82, 4.70, 3.26, 1.66 ppm; 6g,7.16 (2H, dd, J = 5 . 7 , 8.5 Hz); 6.97 (2H, dd, J = 8.5, 8.7 Hz),6.2 (lH , br s), 1.88 (3H, d, J = 1.3 Hz), 1.81 (3H, d, J = 1.3Hz).

[2-Fluor0-5-(2-methyl- -propenyl)phenyl]boronic cid(3g). A solution of l-(p-fluorophenyl)-2-methyl-l-propene 2.0g, 13.3 mmol) in 40 mL of anhydrous THF was chilled in adry-ice/acetone bath. To it was added sec-BuLi (11.3 mL of a1.3 M solution in cyclohexane, 14.6 mmol, Aldrich) over 5 min.The reaction was worked up as described for method B to yieldcrystalline material which was filtered and rinsed with coldether to yield a white solid (2.01 g, 93%). This material wassufficiently pure for the next step.

Potassium [2-Fluoro-5-(2-methyl-l-propenyl)phenyl]-trifluoroborate (2g). To a solution of [2-fluoro-5-(2-methyl-1-propenyl)phenyllboronic acid (2.01 g, 12.3 mmol) in 15 mLmethanol was added aqueous KHFz (12.3 mL of a 3.0 Maqueous solution, 36.9 mmol). After st irring for 1 h, thesolvent was removed (aspirator) to yield a solid which wasdissolved in 30 mL of THF and suction filtered. The solventwas removed t o yield a solid (2.55 g, 81%). Pure material (2.4

g, 76% yield from the boronic acid) was obtained by crystal-lization from tetrahydrofuran, mp 193- 193 C, colorlessplates. Analytical TLC on silica gel, EtOAc, Rf 0.41. Anal.Calcd: C, 46.89; H, 3.94, found: C, 46.48; H, 4.04; IR (KBr,cm-')3021, =C-H; 951, B-F; 200 NMR (CD&N, ppm) 6 7.22(lH , dd, J = 2.5, 5.8 Hz), 6.95 (l H, ddd, J = 2.5, 5.8, 8.2 Hz),6.74 (l H, dd, J = 8.2, 9.0 Hz), 6.18 ( lH , s), 1.79 (3H, d, =1.3 Hz), 1.75 (3H, d, J = 1.3 Hz); B NMR (160 MHz, CDsCN,ppm) 6 3.31 (9, J = 49 Hz).

Potassium (2,6-Dichlorophenyl)trifluoroborate 2h).Method B was modified to follow ref 14 for metalation of 6h(1.07 g) with n-butyllith ium (4.75 mL, 7.27 mmol) at -78 C(dropwise addition over 30 mi d . The resulting slurry wasstir red 45 min, and B(OMe)3 2.5 mL; 22mmol) in THF (5 mL)was added over 2 min. The suspension immediately becameclear. After h, the mixture was allowed to warm and worked

7/26/2019 Vede Js 1995

7/8

3026 J. Org. Chem., Vol. 60, No. 10, 1995

up as usual. The crude 3h15a 1.17g; ca. 2.3:l mixturecontaining the boronic anhydride) was treated with saturatedaqueous KHF2 (5 mL) at room temperature . The thick whiteprecipitate was collected by suction filtration, and the dry solidwas extracted with hot THF. Pure material was obtained byrecrystallization from acetonitrile, 0.976 g (53%) mp 211 Cdec; anal. calcd: C, 28.49; H, 1.20, found: C, 28.69; H, 1.17;200 MHz NMR (CD&N, ppm) 6 7.14 (2H, J = 7.7 Hz, AB2pattern), 6.99 (lH, J = 7.7 Hz, A B 2 pattern); 160 MHz IlBNMR (CDsCN, ppm) 6 2.82 (q, J = 48 Hz).

Potassium 2-Furyltrifluoroborate 2i). Furan (Aldrich,

dried over 3 molecular sieves, 5.0 mL, 68.7 mmol) wasdissolved in 50 mL of anhydrous THF and treated with 42.0mL of a-BuLi (Aldrich, 1.64 M, 68.9 mmol). After stirr ing at-5 C for 3.5 h, th e furyllith ium was treated with B(i-OPr13as described for [3,4-bis(trifluoromethyl)phenylllithium. hecrude boronic acid, obtained after the usual workup (methodB), was dissolved in 200 mL of MeOH and 40 mL uf water,and then 3 equiv of KHF2 (Aldrich, 16 g, 206 mmol) was added.The solution was refluxed overnight. After th e same workupas described for potassium [4-(trifluoromethyl)phenyl]trifluo-roborate, 5.8 g (48%) of product was obtained a s yellowcrystals. TLC analysis on silica gel indicated tha t thiscompound was readily hydrolyzed t o 2-furylboronic acid in wetsolvent (e.g.CH3CN). Pure materia l (5.8 g, 48% from furan)was obtained by recrystallization from anhydrous acetonitrile/ethyl acetate, mp 200 C dec. Anal. Calcd: C, 27.61%; H,1.74%, ound: C, 27.31%; H, 1.33%; IR (KBr, cm-') 1575, C-C;1005, B-F; 970, B-F; 200 MHz NMR (CD&N, ppm) 6 7.44-7.33 (l H, m), 6.25-6.17 (l H, m), 6.17-6.10 (l H, m); 160MHzIlB NMR (CDsCN, ppm) 6 1.8 (9, J = 49 Hz).

p-Fluorophenyl -)-Menthyl ther 6j). The procedureis an adaptation of the method of Whitesides et al.17 To asolution of (-)-menthol (10.3 g, 66 mmol, 1.3 eq, Aldrich) in75 mL of dry THF chilled in an ice-water bath under nitrogenwas added n-BuLi (45 mL of a 1.6 M solution in hexanes, 73mmol, 1.5 eq, Aldrich) over 20 min. The resulting alkoxidewas cannula transferred to a 1 L flask equipped with a refluxcondenser charged with anhydrous CuCl(6.6 g, 66 mmol, 1.3eq, Mallinckrodt). To the resulting dark green solution wereadded anhydrous pyridine (400 mL) and p-bromofluorobenzene(5 .5 mL, 50 mmol, Aldrich). The solution was refluxed for 72h and allowed to cool to room temperature. The reaction wasquenched by dropwise addition of 100 mL of 10% aqueous HCl.

The mixture was ether extracted (4 x 200 mL), and thecombined extracts were washed with distilled water (2 x 100mL) and brine (2 x 100 mL). Removal of solvent (aspirator)yielded 14.0 g of crude solid, which was fractionally sublimedt o yield (-)-menthol (55 C, 0.1 torr, 4.1 g, 43% recovery) intwo fractions and pure p-fluorophenylmenthyl ether (75-80

C, 0.1 torr, 8.9 g, 71%) in the remaining three fractions.Analytical TLC on silica gel, 1:4 etherhexane, R f = 0.63. Purematerial was obtained by sublimation (75-80 C, 0.1 torr), mp47.5-48 C, colorless needles; m l e , calcd for C1&&0 250.1733;found 250.1720, 1 ppm error; IR (KBr, cm-') 2967, =C-H;1208, C-0; 500 MHz NMR (CDC13, ppm) 6 6.95 (2H, dd, J =8.2, 9.5 Hz), 6.83 (2H, dd, J = 4.6, 9.5 Hz), 3.91 ( lH, ddd, J =4.1,10.5,10.5Hz), 2.22 (l H, d sept, J = 2.8, 7.0Hz) , 2.10(1H,dddd, J = 1.9, 3.8, 3.8, 10.9 Hz), 1.73-1.68 (2H, m), 1.48 ( lH,dddd, J = 2.8, 2.8, 10.5, 12.5 Hz), 1.46-1.38 ( lH, m), 1.13-1.04 (l H, m), 1.03-0.95 (l H, m), 0.92 (3H, d, J = 7.0 Hz), 0.91(3H, d, J = 7.0 Hz), 0.95-0.91 ( lH, m), 0.78 (3H, d, J = 7.0Hz).

Potassium [2-Fluoro-6-((-)-menthyloxy)phenylltriflu-oroborate 2j). A solution of p-fluorophenyl (-)-menthylether (0.67 g, 2.7 mmol) in 40 mL of anhydrous THF waschilled in a dry ice/acetone bath. To it was added s-BuLi (3.2mL of a 1.0 M solution in cyclohexane, 3.2 mmol, Aldrich) over10 min. The resulting yellow solution was stirred for 70 minand triisopropyl borate (0.74 mL, 3.2 mmol, distilled oversodium, Aldrich) was added in one portion. The mixture wasallowed to warm to room temperature and was quenched bythe addition of 25 mL of 10% aqueous HC1, followed by dilutionwith 20 mL of ether. The organic layer was extracted with 1N NaOH (3 x 50 mL), and the combined basic extracts wereacidified to pH = 3 with 10% aqueous HC1. The mixture was

Vedejs et al.

ether extracted (3 x 50 mL) and the organic portion was driedover Nap904 and concentrated (aspirator) o yield a colorlessoil weighing 0.67 g in 85 crude yield. The oil was dissolvedin 25 mL of methanol, and to the solution were added KHFz(0.36 g, 4.6 mmol, Aldrich) and 5 mL of distilled water. Thesuspension was refluxed for 24 h and cooled to room temper-ature. The solvent was removed in vacuo to yield a white solidwhich was dissolved in 10 mL of hot acetonitrile and hotfiltered. The solvent was removed t o yield 0.64 g of crystallinematerial in 78% yield. Analytical TLC on silica gel, EtOAc,R f = 0.46. Pure material (0.62 g, 76%) was obtained by

crystallization from eth erh exa ne, mp 146-147 C, colorlessneedles. Anal. Calcd: C, 53.93; H, 6.24, found: C, 53.36; H,6.39; IR (KBr, cm-') 2956,=C-H; 2870, C-H; 986, B-F; 270MHz NMR (CDC13, ppm) 6 6.93 ( lH, d, J = 4.4 Hz), 6.57 (2H,

( lH, m), 2.1-1.9 (2H, m), 1.62 (2H, d, J = 9.4 Hz), 1.42-1.30(2H, m), 1.25-0.95 (2H, m), 0.85 (3H, d, J = 6.9 Hz), 0.80 (3H,d, J = 6.4 Hz), 0.69 (3H, d, J = 6.9 Hz); IlB NMR (160 MHz,CD3CN, ppm) 6 3.19 (9, J = 44 Hz).

(2R,4R)-3-[ Dimethylamino)methylidene]-2-fluoro-2-phenyl-1,3,2-oxazaborolidin-S-one 7-trans). &Phenyl-glycine (Aldrich; 0.914 g, 6.05 mmol) was dissolved in 1 equivmethanolic of NaOMe (8.2 mL, 0.74 M, 6.1 mmol; preparedfrom Mg-dried methanol and sodium) at room temperatureunder a nitrogen atmosphere. Dimethylformamide dimethylacetal (Aldrich; distilled a t 1 atm, bp 102-4 C;0.762 g, 6.4mmol) was added, and the solution was stirred for 75 min.Concentration to a white foam (rotary evaporator, 40 C)followed by triturat ion with CH2Clfit20 and drying (0.5 mm,40 C, 12 h) afforded a white solid, Me2NCH=NCH(Ph)C02-Na, used without further purification. The crude dry salt (1.38g, 6.05 mmol) and potassium phenyltrifluoroborate (1.15 g,6.25 mmol) were suspended in 80 mL anhydrous THF undernitrogen at room temperature and treated with 2.3 equiv ofchlorotrimethylsilane (1.75 mL, distilled from polyvinylpyri-dine) in one portion. After st irring for 2 h, the volatiles wereremoved by bulb-to-bulb distillation at room temperatureunder static vacuum (ca. 0.5mmHg). The resulting white solidwas dissolved in 3:l CH2Clfi20 (20 mL), the aqueous layerwas washed with additional CHzClz (5 mL) and the combinedorganic extracts were washed with saturated aqueous NaCl,dried (Na2SOdjMgSO4), and concentrated to a foam (rotaryevaporator, 25 C). The crude residue after solvent removal

at 0.5 mm was washed with water (15 mL) and ether (15 mL)and then dissolved in CHzCl2 (400 mL), dried (Na2SO&IgSO4),and concentrated t o a solid (rotary evaporator, 30 C). Thecrude product (1.60 g, 85 ) consisted of a 99:l mixture ofdiastereomers 7-trans:7-cis, s determined by analyticalHPLC [5 pm silica gel, 250 mm x 4.6 nun, 25% ethanolhexaneeluent 1.5 mum in, t~ = 7.5 min (major) and t~ = 10.50 min(minor)]. Crystallization from anhydrous CH2Clfit20 at roomtemperature produced 1.47 g (three crops, 78% based onsta rt ing D-phenylglycine) of pure diastereomer 7-trans; na-lytical TLC on silica gel, 2:l EtOAchexane, Rf = 0.26;recrystallization from etheddichloromethane, mp 217-219 Cdec; C17HlsBFN202; m l e 312.1433; base peak = 235 amu; IR(CH2C12, cm-') 1755, C-0; 1680, C=N; 200 MHz NMR (CD3-CN, ppm; amidine EIZ rotamers) 6 7.60-7.22 ( l lH , m) 5.62

(0.75H, s) 2.87 (2.25H, s) 2.78 (0.75H, s) 2.78 (2.25H, s); IIBNMR (proton decoupled, CH2C12, ppm) 6 7.15; 19F NMR (470MHz, ppm vs CFaCsH5 = 0 ppm), 6 -86 ( E , major), -91 2 ,minor). The labile (minor) diastereomer 7-cis could not beobtained pure; analytical TLC on silica gel, 2: l EtOAchexane,R f = 0.18; 200 MHz NMR (CDsCN, ppm, partial) 6 5.66 (0.4H,s) 5.28 (0.6H, s) 2.99 (1.8H, s) 2.94 (1.2H, s) 2.83 (1.8H, s) 2.71(1.2H, s); 19F NMR (470 MHz, ppm vs CF3CsH5 = 0 ppm), 6-83 ( E , major), -89 2 , minor).

(2R,4R)-3-[ Dimethylamino)methylidenel-2-fluoro-2-(2-fluorophenyl)-1,3,2-oxazaborolidin-6-one &trans). Toa suspension of sodium (R)-N-R(N'JV'-dimethylamino)meth-ylidenelphenyllglycine (0.44 g, 1.98 mmol) in THF (30 mL) wasadded chlorotrimethylsilane (0.50 mL, 3.96 mmol, freshlydistilled over CaHz and stored over polyvinylpyridine, Aldrich).The suspension cleared to a solution and then became cloudy

d, J = 7.4 Hz), 3.84 ( lH , ddd, J = 3.7, 6.6, 6.6 Hz), 2.3-2.1

(0.25H, d, 4Jm = 3.7 Hz) 5.33 (0.75H, d, 4Jm = 3.7 Hz) 2.89

7/26/2019 Vede Js 1995

8/8

Synthesis of Potas sium Aryltrifluoroborates

over time. The mixture was stirred for 20 min and was chilledin a dry ice/acetone bath . A suspension of potassium (2-fluoropheny1)trifluoroborate (0.40 g, 1.98 mmol) in 20 mL ofTHF was added over 20 min. The mixture was stirred at -78C for 1 h and was allowed to warm to room temperature. ARer

2 h, ethyl acetate (30 mL) was added, and the mixture wasextracted with saturated aqueous NaHC03 (20 mL), distilledwater (2 x 20 mL), and brine (2 x 20 mL). The organic layerwas dried over NaZS04, and the solvent was removed (aspira-tor) t o yield 8-trans s a hygroscopic white solid. A solutionenriched in the minor isomer 8 4 s ca. : l mixture) was storedin dichloromethane-ether to give colorless crystals, mp 207-208 C; analytical TLC on silica gel, 1: l EtOAchexane, Rf =0.52; 300 MHz NMR (CD2C12, 1:l rotamer mixture, ppm) 67.58 (0.5H, s), 7.53-7.43 (0.5H, m), 7.42-7.33 (3.5H, m), 7.27-7.19 (2.5H, m), 7.10-7.05 (1.5H, m), 7.02-6.84 (1.5H, m), 5.5(0.5H, s), 5.14 (0.5H, s), 3.03 (1.5H, s), 2.86 (1.5H, s), 2.85( M H , ), 2.72 (1.5H, s), 1.75 (br s, HzO). Traces of the signalsof 8-trans were also present (ca. 5%). I1B NMR (160 MHz, vsBF3eEt20 external reference, CDC13, ppm) 6 6.95, 5.95 (tworotamers). Attempted recrystallization did not remove thecontaminant 8-trans, nd decomposition was detected. How-ever, recrystallization of 8-trans ave a single isomer; analyti-cal TLC on silica gel, dichloromethane, Rf = 0.25; colorlessprisms (0.46 g, 71%), hygroscopic; mp 212-213 C from O : lethyl acetate:acetonitrile; IR (KBr, cm-l) 1741, C=O; B-0;300 MHz NMR (CDC13, ppm) 6 7.72 (0.2H, ddd, J = 1.9, 7.0,

7.0 H z), 7.70 (0.8H, ddd, J = 1.9, 7.0, 7.0 Hz), 7.55-7.12 (8H,m), 6.94 (0.2H, dd, J = 8.2, 9.7 Hz), 6.89 (0.8H, dd, J = 8.2,9.7 Hz), 5.44 (0.2H, d, J = 3.5 Hz), 5.20 (0.8H, d, J = 3.5 Hz),3.01 (0.6H, s), 2.99 (2.4H, s), 2.94 (2.4H, s), 2.85 (0.6H, s), 1.60(br s, HzO); llB NMR (160 MHz, BF3-etherate as ex ternalreference) 6 6.06 ppm.

34 Dimethylamino)methylidene]-2-fluoro-2-[4-(triflu-oromethyl)phenyll-1,3,2-oxazaborolidin-5-one 9). Thesodium sal t of N-[(N',N-dimethylamino)methylidene]gly~ine~~(150 mg, 1.0 mmol) and potassium [4-(trifluoromethyl)phenyll-trifluoroborate (250 mg, 1.0 mmol) were suspended in 20 mbanhydrous acetonitrile (distilled from CaH2, stored over 3 Amolecular sieves) at room temperature under nitrogen. Tri-ethylamine (Aldrich, distilled from CaH2,0.071 mL, 1.0 mmol)and TMSCl (Petrarch, distilled from CaH2, 0.38 mL, 0.10mmol) were added. After 3 h stir ring, more Et3N (0.1 mmol)and TMSCl (0.05 mmol) were added and sti rring was contin-ued 2 h. The mixture was poured into rapidly stirred ice-coldphosphate buffer (pH 7) and ethyl ace tate (40 mL). Theorganic layer was separated, washed with water ( 3 x 20 mL)and brine (20 mL), dried (MgS04), and evaporated (aspirator) .The residual oil was crystallized from CHzClz (25 mL) andether (30 mL) to give 236 mg of white crystals (first crop) + 4mg (second crop), 79% combined yield. Analytical TLC onsilica gel, EtOAc, R f = 0.10, mp 164.0-165.5 C. Formula,C12H13BFaz02; ml e, M 1,305.1094; error = 3 ppm; IR (KBr,cm-l) 1753, C=O; 1741, C=O; 1681, C=N; 270 MHz NMR(CDC13, ppm; 2:l mixture of amidine rotamers) 6 7.60 (1.33H,d, J = 8.1 Hz), 7.56 (1.33H, d, J = 8.1 Hz), 7.52 (1.33H, s),7.51 (0.33H, br s), 6.96 (0.67H, br s), 4.51 (0.67H, d, J = 17.2Hz), 4.36 (0.67H, d, J = 17.2 Hz), 4.35 (0.67H, s), 3.27 (2H, s),3.09 (3H, SI 2.86 (l H, d, J = 1.0 Hz).

34 Dimethylamino)methylidenel-2-fluoro-2-(2-furyl)-1,3,2-oxazaborolidin-5-one 10). The usual procedure wasused for the preparation of the title compound. Thus, thesodium salt of N-[(N,N-dimethy1amino)methylidenelglycine(343 mg, 2.25 mmol) and potassium 2-furyltrifluoroborate (2i)(407 mg, 2.34 mmol) were suspended in 25 mL anhydrousacetonitrile under nitrogen and treated with Et3N (0.35 mL,2.5 mmol) and TMSCl (0.61 mL, 2.5 mmol). After sti rring atroom temperature for 4 h, all volatiles were removed byaspirator, and the oily residue was taken up with 20 mL ofCHzClz, washed with ice-water (3 x 15 mL), and dried(MgS04). Half of the above solution was concentrated to 2 mL,and anhydrous ether was added until cloudiness persisted.Crystallization gave 83 mg of yellow blocks. The mother liquorwas concentrated and treated with ether in the same manner,yielding another 15 mg of yellow crystals, combined yield 38%.The amidine group existed as a 2:l mixture of rotamers in

J. Org. Chem., Vol. 60, No. 10, 1995 3027

CDC13. Analytical TLC on silica gel, EtOAc, Rf = 0.12.Analytical mate rial was obtained by filtration through a silicagel plug (0.5 x 2 cm) and subsequent crystallization from ethylacetatel ether, mp 103.8-104.2 C, yellow crystals. Molecularion calcd for CsHlzBFNz03: 226.09240; found mle = 226.0928,error = 1 ppm; IR (KBr, cm-l) 1747, C=O; 1740, C=O; 1690,C=N; 200 MHz NMR (CDsCN, ppm) 6 7.55-7.45 (1.33H, brm), 7.17 (0.67H, br s), 6.43 (0.67H, dd, J = 3.1, 0.7 Hz), 6.37(0.33H, dd, J = 3.1,0.7 Hz), 6.35-6.30 (l H, br m), 4.45 (0.67H,d, J = 17.3Hz),4.35(0.67H,dd, = 17.3,2.8Hz),4.18(0.67H,br s), 3.20 (2H, s), 3.06 (2H, s), 3.04 (lH, s), 2.92 (l H, d, J =

1.3 Hz).l-Phenyl-l,3-pentanedione om pl ex 11. Potassium2-(fluorophenyl)trifluoroborate (2e) 0.20 g, 1.0 mmol) andl-phenyl-l,3-pentanedione 0.17 g, 1.0 mmol) were placed ina 10 mL flask, and the system was purged with nitrogen. Dryacetonitrile (7.0 mL) and chlorotrimethylsilane (0.25 mL, 2.0mmol, freshly distilled over CaH2 and stored over polyvinyl-pyridine, Aldrich) were added sequentially. The mixture wasstirred 1 h at room temperature and then diluted with 10 mLof ethyl acetate, washed with distilled water (2 x 20 mL) andbrine (2 x 20 mL), and dried over Na2S04. Removal of solvent(aspirator) gave an oily solid which was subjected t o plugfiltration on silica (3 x 4 cm) with CHzClz as eluent, yieldingafter removal of solvent (aspirator) a white solid (0.297 g, 99%).Analytical TLC on silica gel, EtOAc, R f = 0.71. Pure materialwas obtained by crystallization from etherhexane, mp 69-70

C, colorless needles. Molecular ion calcd for C17H15BF202:300.11328; found ml e = 300.1139, error = 2 ppm; base peak= 205 amu; IR (KBr, cm-l) 1611, C=O; B-0; 2986, C-H; 300MHz NMR (CDC13, ppm) 6 8.2-8.0 (2H, m), 7.68 (lH, ddd, J= 7.4, 7.4, 1.9 Hz), 7.63 ( lH , dddd, J = 7.4, 7.4, 1.2, 1.2 Hz),7.4-7.2 (2H, m), 7.28 (l H, dddd, J = 7.4, 7.4, 5.8, 1.9 Hz),7.13 ( lH , dddd, J = 7.4, 7.4, 1.2, 0.8 Hz), 6.95 (l H, ddd, J =9.4, 7.4, 0.8 Hz), 6.55 (l H, s), 2.63 (2H, q, J = 7.8 Hz), 1.27(3H, t, J = 7.8 Hz); llB NMR (160 MHz, BFJetherate, CD3-CN, ppm) 6 5.94 (s).

Complex 12 from 2j an d l-Phenyl-1,3-pentanedione.The chiral aryltrifluoroborate 2j (0.36 g, 0.1 mmol) andl-phenyl-1,3-pentanedione 0.018 g, 0.1 mmol) were combinedin a 10 mL flask. The system was flushed with dry nitrogen,and dry acetonitrile was added (5 mL). To the resultingsolution was added chlorotrimethylsilane (26 pL, 0.2 mmol)and a white precipitate immediately formed. The suspensionwas stir red for 30 min and was quenched with distilled water.The mixture was diluted with 10 mL of ethyl acetate andextracted with distilled water (2 x 20 mL) and brine (2 x 20mL) and dried over Na2S04. After removal of solvent (aspira -tor), the residue was purified by preparative layer silica gel(20 x 20 x 0.1 cm), with dichloromethane eluent , to give 0.040g of 12 n 89% yield as a l ight yellow oil; analytical TLC onsilica gel, dichloromethane, Rf = 0.81. Molecular ion calcd forC27H33BF203: 454.24905; found m le = 454.2502, error = 2ppm; M 19,435.2495, error = 3 ppm; base peak = 205 amu;IR (neat, cm-l) 2869, C-H; 1541, C=O; 300 MHz NMR (CDCl3,ppm) 6 8.05 (2H, d, J = 7.8 Hz), 7.65 (l H, dd, J = 7.4, 7.4 Hz),7.5 (2H, dd, J = 8.2, 7.4 Hz), 7.2 (l H, dd, J = 4.7, 3.1 Hz),6.85 ( lH , dd, J = 8.6, 8.6 Hz), 6.78 (l H, ddd, J = 8.6, 4.7, 3.1Hz), 6.56 (lH , s), 3.98 ( lH , ddd, J = 10.5, 10.5, 3.9 Hz), 2.65(2H, q, J = 7.4 Hz), 2.3-2.1 (2H, m), 1.72-1.44 (4H, m), 1.31(3H, t, J = 7.4 Hz), 1.31-0.92 (3H, m), 0.94 (3H, d, J = 7.0Hz), 0.91 (3H, d, J = 7.0 Hz), 0.80 (3H, d, J = 7.0 Hz); llBNMR (160 MHz, BFJetherate, CD3CN, ppm) 6 5.90.

Acknowledgment. This work was supported by theNational Institutes of Health GM44724 n d instru-mentation grant 1 S10 RR08389-01).

Supplementary Material Available: Copies of lH NMRspectra of 6g, j, 7-trans-E, -trans-E, nd 9-12 8 pages).This material is contained in libraries on microfiche, im-mediately follows thi s article in the microfilm version of thejournal, and can be ordered from the ACS; see any currentmasthead page for ordering information.

J0941646A